Using a neural network control policy for rapid switching between beam parameters in an FEL

A. L. Edelen¹, S. G. Biedron², J. P. Edelen³, S. V. Milton⁴, P. J. M. van der Slot⁵

¹Colorado State University, ²Element Aero, ³Fermi National Accelerator Laboratory, ⁴Los Alamos National Laboratory, ⁵University of Twente

38th International Free Electron Laser Conference

Abstract

FEL user facilities often must accommodate requests for a variety of beam parameters. This usually requires skilled operators to tune the machine, reducing the amount of available time for users. In principle, a neural network control policy that is trained on a broad range of operating states could be used to quickly switch between these requests without substantial need for human inter-vention. We present preliminary results from an ongoing study in which a neural network control policy is investi-gated for rapid switching between beam parameters in a compact THz FEL.

Read the paper: http://jacow.org/fel2017/papers/wep031.pdf

Contact: Auralee Edelen